OBJECTIVE: To predict effects over time on cardiometabolic risk factors of adding rimonabant to diet and exercise in overweight/obese patients. Both weight-dependent and weight-independent effects of treatment were examined in order to carry out a Discrete Event Simulation (DES). The DES predicts the time-dependent course of individual patients' data according to treatment. METHODS: Data were taken from 4 RIO trials in over 6600 overweight/obese patients who received once-daily rimonabant 20mg (or placebo) on top of diet and exercise. Time-dependent functions of individual changes in weight, waist circumference, cholesterol, HDL-cholesterol, triglycerides, HbA1c, fasting glucose, and blood pressure were required for the DES. Change over time was analyzed in pooled trial data using a two-step process: 1) Logistic regressions predicted whether the parameter would decrease, and 2) fractional polynomials predicted change as a function of time and other factors. Random-effects were included to account for within and between patient variance. RESULTS: One year of treatment induced cardiometabolic improvements following a curvilinear course with steep initial changes and subsequent stabilization after six to nine months. To properly reflect these time-dependent changes, several time parameters were required in each equation. The degree of change over time depended on baseline levels and other patient characteristics and on weight changes, but rimonabant 20 mg provided an additional, weight-independent, statistically-significant, effect (OR for worsening 0.36-0.88 depending on outcome considered). CONCLUSIONS: Multivariate time-dependent analytic techniques can provide the detailed estimates required for discrete event simulation. Growth curves enable estimation of the course of each simulated individual over time by providing for realistic modeling of risks and facilitate probabilistic sensitivity analyses. These analyses provided a detailed, accurate reflection of rimonabant's effects, showing that its beneficial impact is present even when accounting for any reductions in weight, suggesting that there is an effect beyond weight loss.